Tuesday, February 7, 2012

New Perspectives on Lake Effect

Skiers love the Great Salt Lake effect, but it is a serious thorn in the side of Utah weather forecasters.  So much so that we call it the "dreaded lake effect" or DLE for short.  

We've just submitted a paper entitled Great Salt Lake-Effect Precipitation: Observed Frequency, Characteristics, and Associated Environmental Factors that presents a revised climatology that we hope will improve both understanding and prediction.  The project was spearheaded by Trevor Alcott, a graduate student here at the University of Utah, and includes collaboration with Neil Laird, a professor at Hobart and William Smith Colleges in upstate New York.  Neil's students Benjamin Albright and Jessica Popp performed the initial event identification for the climatology.  

The climatology is based on 149 lake-effect events that were identified over 13 seasons.  Skiers will be interested to learn that the event frequency is greatest in the fall (Oct–Nov) and spring (early Apr) with a mid-winter minimum.  This contrasts with the mid-winter maximum identified in an earlier study by my group (Steenburgh et al. 2000), which was based on a much shorter period of record and a small number of events.  

Number of lake-effect events by half month
(Alcott et al. 2012)
One of the more interesting aspects of the Great Salt Lake effect is its diurnal (i.e., day to night) modulation.  Events tend to trigger after sunset and dissipate after sunrise.  This diurnal modulation is strongest in the spring (Mar–May) and weakest in the winter (Dec–Feb).   

Number of days with lake effect by time of day [UTC
and Local Standard Time indicated (Alcott et al. 2012)]
For skiers looking for bluebird days, this is a great result.  Snowfall at night, clearing skies during the day.  Of course, as we will discuss in a few weeks when we submit a second paper on this subject, the amount of snow produced by lake-effect is nowhere near as large or as important as suggested by conventional wisdom, hype, and ski brochures.  

As many people recognize, the movement of a cold airmass over the relatively warm lake surface is a necessary but not sufficient condition for lake-effect.  One of the more interesting findings of the study is that lake-effect events during winter more frequently occur during periods of weaker differences in temperature between the lake and the atmosphere at 700 mb (10,000 ft) (a.k.a. the lake-700 mb temperature difference, which is commonly used to assess instability over the lake).   Or, alternatively, lake-effect in the spring and fall typically requires a larger difference in temperature between the lake and the atmosphere at 700 mb.  

The paper identifies some of the shortcomings of existing forecast techniques, which are based primarily on the difference between the lake temperature and the 700-mb temperature.  Instead, we propose a new approach based on a seasonally varying lake-700 mb temperature difference threshold (which we call ΔTexcess) and low-level relative humidity.  The historical likelihood of lake-effect based on these two variables is shown below.    

Fraction (%) of soundings with lake effect as a
function of ΔTexcess and low-level relative humidity
(Alcott et al. 2012)
In the paper, we also discuss how wind direction and mid-level relative humidity affect the likelihood and coverage of lake-effect precipitation.  

These results will hopefully improve lake-effect forecasting.  I suspect that the primary benefit will be a reduction of false alarms, especially events where lake-effect is forecast but doesn't occur.  That being said, this study is not a panacea.  We have more work to do and I suspect forecasters will continue to call it the dreaded lake effect (or something saltier) in the coming years. 

Stay tuned for a second paper on lake-effect that should prove even more interesting for skiers.  


  1. how much does wind speed, if at all, play into such events? I am from NE Ohio and I remember growing up that it was always a fairly nice day but windy when we would get lake effect squalls. In between squalls would be perfectly blue skies too.

  2. The fall maximum is probably pretty universal in regard to lake effect, but I suspect the spring maximum is fairly unique to the GSL... do you know if this is the case? I think a lot of it has to do with how shallow the lake is, allowing it to warm dramatically during even brief periods of warm/sunny spring weather. Plus the related fact that our ridge vs. trough temperature variance is so large here in the spring.

  3. WIth regards to wind speed, things work a bit differently here. Strong wind events in the Great Lakes can produce multiple narrow snowbands that are aligned along the wind and separated by cloud free zones. I can't recall ever seeing such an event on the Great Salt Lake. Instead, stronger winds show some *tendency* to produce isolated bands here (rather than multiple bands). This is discussed in the paper, although admittedly in a techy fashion.

    With regards to the spring maximum, I don't know if this has been documented elsewhere. It could be very unique to the Great Salt Lake. It is true that the shallow (mean depth 3 meters) nature of the lake helps it to warm fast, but the hypersalinity also means that the lake doesn't freeze. Few inland water bodies have such characteristics.

    The fall maximum for the GSL is probably a bit earlier than over the Great Lakes. Boonville, NY, for example, gets fairly similar amounts of precipitation from Nov-Jan. Similarly, on Hokkaido Island (Japan), the peak is in Jan. The shallowness of the GSL probably works against it in this regard.

  4. Any signs of the storm door opening in the medium range?

  5. Maybe something early next week. Too much spread amongst the various model forecasts, however, to get excited. Give it a couple of days and we'll see how things look.

  6. Great post! I like the science-made-relevant aspect...

  7. Thanks, Jim. This is exactly the sort of thing that is applicable to our operation at the DOT. Another excellent example of bridging the research-operations gap.